The present invention relates to an air-fuel ratio control system for an engine of a motor vehicle, and more particularly to a system having an electronic fuel injection system controlled by learning control.
In one type of electronic fuel-injection control, the quantity of fuel to be injected into the engine is determined in accordance with engine operating variables such as mass air flow, intake-air pressure, engine load and engine speed. The quantity of fuel is determined by a fuel injector energization time (injection pulse width).
Generally, a desired injection amount is obtained by correcting a basic quantity of injection with various correction or compensation coefficients of engine operating variables. The basic injection pulse width is derived from a lookup table to provide a desired (stoichiometric) air-fuel ratio according to mass air flow or intake-air pressure and engine speed. The basic injection pulse width T.sub.P is expressed, for example, as follows. EQU T.sub.P =f ( P, N )
where P is intake-air pressure and N is engine speed.
Desired injection pulse width (T) is obtained by correcting the basic injection pulse T.sub.P with coefficients for engine operating variables. The following is an example of an equation for computing the actual injection pulse width. EQU T=T.sub.P .times.K.times..alpha..times.Ka
where K is at least one of coefficient selected from various coefficients such as coefficients on coolant temperature, full throttle, etc., .alpha. is a feedback correcting coefficient which is obtained from output signal of an O.sub.2 -sensor provided in an exhaust passage, and Ka is a correcting coefficient by learning (hereinafter called learning control coefficient) for compensating the change of characteristics of devices with time in the fuel control system such as, injectors and an intake air pressure sensor, due to deterioration thereof. The coefficients K and Ka are stored in lookup tables and derived from the tables in accordance with sensed informations.
The control system compares the output signal of the O.sub.2 -sensor with a reference value corresponding to desired air-fuel ratio and determines the feedback coefficient .alpha. so as to converge air-fuel ratio of air-fuel mixture to the desired air-fuel ratio.
As described above, the basic injection pulse width T.sub.P is determined by the intake-air pressure P and engine speed N. However, the intake-air pressure is not always constant, even if the engine speed is the same as previous speed. For example, when a valve clearance (the clearance between an intake (or exhaust) valve-stem tip and a rocker arm) becomes large with time, the valve opening time becomes short. As a result, overlapping times of the intake valve opening time and the exhaust valve opening time become short. Accordingly, quantity of exhaust gas inducted into an intake passage from a combustion chamber during the overlapping time becomes small. Thus, quantity of the intake-air increases. However, the intake-air pressure and hence quantity of fuel injection do not change. Accordingly, the air-fuel ratio becomes large (lean air-fuel mixture). The same result occurs when driving at high altitude.
Such a change of characteristic of a device is also corrected by updating a learning control coefficient. In a prior art, for example U.S. Pat. No. 4,445,481, the learning control coefficient is updated little by little. Accordingly, it takes long time to get a desired coefficient, which causes the delay of control of air-fuel ratio.